Data from: Hydromorphological restoration of streams does not ensure the recovery of ecosystem functioning
Data files
Sep 17, 2026 version files 197.88 KB
-
Dataset_Haase_et_al.csv
177.77 KB
-
README.md
5.26 KB
-
References_Haase_et_al.csv
14.85 KB
Abstract
Ecological restoration is widely expected to rebuild biodiversity and ecosystem functioning simultaneously. Yet ecosystem functions emerge from multiple environmental controls operating across spatial and temporal scales, raising the question of whether restoring habitat structure and biological communities is sufficient to recover ecosystem functioning. We compiled 582 effect sizes from 66 studies, covering 19 biodiversity attributes and 9 ecosystem functions, to quantify how ecosystem functions responded to hydromorphological degradation and restoration and assessed whether biodiversity recovery consistently coincided with ecosystem function recovery. Functional responses varied markedly among ecosystem processes and between benchmarks (reference/ impacted), with no universal response across restoration strategies. While gross primary production and organic matter decomposition exceeded reference conditions in degraded streams and restored streams alike, nitrate uptake remained lower than in reference streams after restoration. Denitrification responses were significantly lower in restored streams relative to degraded reaches. Differences in responses between biodiversity and ecosystem functioning varied across the studied functions and did not change systematically with restoration age. Biodiversity–ecosystem functioning relationships were similarly inconsistent across functions and associated with high uncertainty. Together, our results indicate that ecosystem functioning does not recover as a single integrated property following hydromorphological restoration. Instead, individual ecosystem processes appear to recover along partially independent trajectories determined by distinct environmental controls, including nutrient supply, hydrologic exchange, organic matter dynamics, thermal regimes, and catchment-scale stressors. Functional restoration, therefore, requires more than rebuilding habitat structure and biodiversity alone: it requires restoring the environmental process domains under which ecosystem functions emerge. This shifts restoration from a paradigm centered on structural habitat recovery toward one explicitly grounded in process-based ecosystem recovery.
Dataset DOI: 10.5061/dryad.m37pvmdjq
Description of the data and file structure
This is the accompanying dataset and literature references used for the meta-analysis presented in Haase et al. (2026): Hydromorphological restoration of streams does not ensure the recovery of ecosystem functioning.
Haase et al. quantified how ecosystem functions responded to hydromorphological degradation and restoration and assessed whether biodiversity recovery consistently coincided with ecosystem function recovery.
Files and variables
File: References_Haase_et_al.csv
Description: The file contains literature references from which the data were extracted. Studies and extracted data (Dataset_Haase_et_al.csv) are connected via study.id.
File: Dataset_Haase_et_al.csv
Description: The dataset comprises the extracted data from studies and calculated log response ratios.
Missing values are blank cells.
Explanations of the column headers in the spreadsheet dataset.
| Header | Description |
|---|---|
| study.id | ID used to code the study |
| ESF.var | Ecosystem function |
| var.cat | Category of variables [structure=biodiversity variables, function=ecosystem function variables] |
| unit | Unit of measurement in which the variable was recorded |
| design.study | Study design [CI_P, Control Impact positiv; CI_N, Control Impact negative, BA Before after] |
| ctr.sitepublished | Name of the control site as extracted from the study; if more than one site was analysed, it was coded with "reference", "degraded" |
| rest.sitepublished | Name of the restored site as extracted from the study; if more than one site was analysed, it was coded with "restored" |
| ctr.lat.num | Geographical latitude of the control site; means were used if there was more than one site |
| rest.lat.num | Geographical latitude of the restored site; means were used if there was more than one site |
| ctr.long.num | Geographical longitude of the control site; means were used if there was more than one site |
| rest.long.num | Geographical longitude of the restored site; means were used if there was more than one site |
| ctr.ORD.STRA | Strahler stream order of control site |
| rest.ORD.STRA | Strahler stream order of restored site |
| rest.age.in.y | Time since restoration [years] |
| restoration.measure | Type of restoration measure |
| ctr.sitetype | Type of control |
| ctr.n | Number of measurements of control site |
| rest.n | Number of measurements of restored site |
| ctr.mean | Mean value of control site |
| rest.mean | Mean value of reference site |
| ctr.sd | Standard deviation of control site |
| rest.sd | Standard deviation of restored site |
| log.rr | Log response ratio |
| var.rr | Variance of log response ratio |
